Tim L.O. DavisGrant J. WestCSIRO Division of Fisheries. Marine Laboratories,GPO Box J538. Hobart. Tasmania 700 1, Australia

Maturation, reproductiveseasonality, fecundity, andspawning frequency in Lutjanusvittus (Quoy and Gaimard) fromthe North West Shelf of Australia

Abstract.-The reproductive bi­ology of Lutjanus vittus, an impor­tant component of the trawl-fisherycatch from the North West Shelf ofAustralia, was determined from ran­dom. stratified trawls made every2 mo during August 1982-0ctober1983. The smallest mature femalewas 142 mmFL; half of the femalesat 154mm (estimated l+yr of age)were mature. The major spawningseason was from September to April,although some spawning occurred inother months. A semilunar patternof spawning activity was observed,with peaks in the proportion ofspawning females -3 d after the newmoon and 6 d after the full moon.Individuals spawn about 22 timeseach lunar month. The diel changesin proportion of fish with ripe 00­

cytes and early- and late-stagepostovulatory follicles, together withdata on maximum oocyte diameter,suggest that most individuals spawnbetween 11:00 and 15:00h on peakspawning days. The relationshipbetween batch fecundity (F) andfish fork length (L in mm) was expo­nential IF=3.656x 10-6 14.093 ) andwith fish weight (W, g) was linearIF=124.2xW-30811. Unlike gonadweight. batch fecundity did notdecline as the spawning seasonprogressed.

Manuscript accepted 17 February 1993.Fishery Bulletin, U.S. 91:224-236 (1993).

224

Lutjanids are of great importance infisheries throughout the tropics andsubtropics, and Lutjanus vittus dis­tributed throughout the Indo-WestPacific is a highly valued componentof the multispecies trawl fishery ofthe North West Shelf of Australia(NW Shelf>, comprising -4% of thetotal catch of -165,000t between1972 and 1980 (Jernakoff & Sains­bury 1990). Its early life historyhas been studied in Japan (Mori1984); its biology, in New Caledonia(Loubens 1980a,b). Information ongrowth and mortality of this specieson the NW Shelf has recently beenpublished (Davis & West 1992). How­ever, there is no published informa­tion on the reproductive biology ofL. vittus that would pertain to themanagement of this fishery on theNWShelf.

Some information on the reproduc­tive biology of over 40 species oflutjanids is available (see review byGrimes 1987). Many studies have in­dicated that lutjanids are multiplespawners, based on the evidence ofmultimodal size-frequency distribu­tions of oocytes, but techniques suchas the hydrated oocyte and post­ovulatory follicle methods (Hunter &Goldberg 1980, Hunter & Macewicz1980), for measuring spawning fre­quency in marine fishes, have notbeen applied to lutjanids (Everson eta1. 1989). In this paper we present

data on oocyte development, size-at­maturity of females, and batch fecun­dity in L. vittus from the NW Shelf.We also demonstrate both diel andsemilunar spawning periodicitythrough an analysis of the temporaldistribution of hydrated oocytes andearly- and late-stage postovulatoryfollicles, and provide estimates ofspawning frequency and annual .fecundity.

Materials and methods

Specimens were obtained from theCSIRO North West Shelf programbetween August 1982 and October1983 (Young & Sainsbury 1985). De­tails of the two-monthly randomstratified trawl surveys of shelf wa­ters within latitudes 116°E and 119°Eare given in Davis & West (1992). Asubsample of 20-40 Lutjanus vittus,which approximately represented thesize-frequency composition of the to­tal catch (Kimura 1977), was selectedfrom each random trawl. Fork lengthwas measured to the nearest 1mmand total weight to the nearest 1g.Fish were sexed and up to 200 ova­ries per cruise were fixed in 10% for­malin in seawater for histological andwhole oocyte examination. Additionalbiological samples were collectedfrom nonrandom trawls to investigatelunar and diel periodicity in spawn-

Davis and West: Reproductive biology of Luljanus vittus from North West Shelf of Australia 225

ing. Samples were also obtained at 4 h intervals dur­ing 6-9 October 1988 to further investigate diel peri­odicity of spawning.

In the laboratory, ovaries were blotted dry and thenweighed to the nearest 0.1 g. Whole oocytes were ob­tained by cutting a complete cross-section 1-2 mm thickfrom the middle of the ovary, and teasing the oocytesapart with needles. Oocytes were examined in waterunder a stereomicroscope using transmitted light andbrightfield illumination. We measured maximum oo­cyte diameter (MOD) as the diameter of the largestoocyte (to the nearest 20 ~m) in the sample. Becauseoocytes were basically spherical. the orientation ofoocytes with respect to the line of measurement wasrandom.

The size-frequency distribution of oocytes within ova­ries at different stages of maturity was determined.Sections 1-2 mm thick were cut from the middle of oneovary, the oocytes teased apart with needles, and fur­ther separated by immersion in an ultrasonic bath for-5 min. Oocytes were pipetted into grooves on a perspexmicroscope slide and examined on a moving stage at­tachment on a stereomicroscope at 50X, using trans­mitted light. Orientation was random. As well as mea­suring each oocyte, we also noted its appearance sothat the stage of development could be assessed.

We used whole-oocyte staging to assess ovarian ma­turity lHilge 1977, Forberg 1983, West 1990). Our

three-stage scheme (Table 1) differed from Hilge's(1977) in that we distinguished late-migratory nucleusoocytes (Fig. Ib) from yolked oocytes (Fig. 1a) by thepresence of some translucence at the periphery ofthe vitelline area (West 1990). Ovaries with late­migratory nucleus oocytes or hydrated oocytes wereconsidered ripe (Fig. leI.

The relative gonadal index (RGI> was used to quan­tify the reproductive cycle (Erickson et al. 1985). Thisindex is independent of body size and is based on themodel, W=(XiS~i, where W is gonad weight (g), S isovary-free body weight (kg), and (Xi and ~i are param­eters estimated for maturity stage i. As ~i did not dif­fer significantly between the three maturity stages(ANCOVA, F=1.304, df 2,1046, p=0.27). a common ~

was used.To study short-term spawning rhythms and to

verify whole-oocyte staging, we histologically examined477 ovaries from one cruise during the height of thespawning season (November-December 1982). Histo­logical sections were 10~m thick and stained withhaematoxylin and eosin. The sequence of oocyte matu­ration is similar in most teleost species (Wallace &Selman 1981). We used the histological criteria ofYamamoto (1956) for assigning oocytes to developmen­tal stages.

The presence of postovulatory follicles was noted inthe histological material. These follicles are readily

Table 1Whole-oocyte staging scheme in Lutjanus vittus based on appearance of the most advanced oocytes (modified from Hilge1977> and corresponding histological stages (Yamamoto 19561 used to describe ovarian maturity.

Whole-oocyte stage

1 UnyolkedTransparent with visible nucleus,becoming granular and translucent withincreasing size.Granular. translucent with brownishtinge. becoming darker with increasingsize, but not opaque.

2 YolkedCompletely opaque except forperivitelline border.

3 RipeDistinguished from Yolked stage when parts ofthe yolk become translucent.

Oocyte is translucent except for the oil droplet.

Histological stage

Perinucleolus stage Cytoplasm purple, outer cellmembrane not differentiated at 400'<.

Yolk-vesicle stage Pale vacuoles (yolk vesicles>and a thin but distinct pink-staining zona radiata bothpresent.

Yolk-granule stage Yolk granules appear as pinkspheres. Fat vesicles coalesce to form larger vesicles aroundthe central nucleus at the end of this stage.

Early-migratory nucleus stage Nucleus moves outfrom center of the cel'l, being replaced by one or two fat vesicles.

Late-migratory nucleus stage Nucleus moves to theperiphery of the cell and some of the yolk granules coalesce.

Hydrated-oocyte stage Yolk granules coalesce to form a uniformpink-staining mass.

226 Fishery Bulletin 91 (2). J993

Figure 2Photomicrographs oftransverse section ofLII.tjallus ['ittlls ovarystained with haema­toxylin and eosin,showing (a) early­stage postovulatoryfollicles, in which thelumen ILl is openand the folds of thefollicle layer (FLI arediscrete. and Ibl late­stage postovulatoryfollicles. in which thefollicle has collapsedand the follicle layeris no longer discrete.FL = follicle layer;L = lumen. Scalebars 10011m.

identifiable within 12 h after spawning because the folds of thefollicle layer are still discrete (Hunter & Goldberg 1980), butthey become progressively more compact until they appear asa solid mass of cells. We recorded an early stage where thelumen was open and the wall of the follicle clearly visible(Fig. 2a) and a late stage where the follicle had collapsed andthe follicle layer could not be traced around its periphery(Fig.2bl.

Batch fecundity was determined using the gravimetricmethod <Hunter et al. 1985) to count numbers of hydrated butunovulated oocytes in weighed subsamples of formalin-fixedovaries. Each subsample consisted of a wedge of tissue extend­ing from the periphery to the lumen of the ovary. In order toestimate the energetic costs of spawning, the dry weight ofhydrated oocytes was determined in nine fish. About 100 00­

cytes were removed from each formalin-fixed ovary. Oocytes

.6

Figure 1Photomicrographs of whole oocytes of Lutjanusvittus representing (a) yolk granule (YGJ stage, inwhich the oocyte is completely opaque except forthe translucent perivitelline border; (b) late-migra­tory nucleus stage, in which parts of the peripheralyolk have coalesced and become translucent(arrowed); and (c) ripe-oocyte stage, in which wholeoocyte is translucent, except for the oil droplet. Scalebars 500l1m.

500~m

a

Davis and West: Reproductive biology of LUfjanus vittus from North West Shelf of Australia 227

were washed in freshwater, counted, dried for at least24 h at 60°C, and weighed to the nearest 0.1 mg.

Size-at-maturity (Lo.50 l was defined as the smallestlength at which half of the fish had yolked or ripe­stage ovaries. Weighted least-squares nonlinear regres­sion was used to select the model (Gompertz curvelthat best described the relationship between propor­tion of mature fish and fish length. The simplest modelwas chosen in which higher parameter models did notsignificantly improve the fit using the F-ratio statistic(Schnute 1981).

RGI's were log-transformed to stabilize variances,and a 2x2 factorial ANOVA was used to test for differ­ences in RGI's between samples collected in Augustand October and between consecutive years. Differ­ences in RGI over the four periods sampled during themain spawning season (September 1982-April 1983lwere tested using ANOVA, and the period effect waspartitioned into a linear component and deviations froma linear component to test for trends.

The log-likelihood ratio X2 was used to test for inde­pendence in contingency-table data (frequency of fishwith and without postovulatory follicles by time of day,month, etc.l. Cells with expected frequencies <5 wereexcluded from statistical tests. Repeated-measuresANOVA was used to test for differences in the num­bers of eggs/mg tissue taken from the inner and outerwalls, and from anterior, middle, and posterior partsof the ovary of five fish. The relationship between logfecundity and log length at different times of the spawn­ing season was examined using ANCOVA.

Results

Maturation of ovaries

Size-frequency and stage of whole oocytes in ovariesrepresenting the developmental sequence of matura­tion (Fig. 3) indicated that the size of oocytes is corre­lated with their stage of development but there is con­siderable overlap between stages. The pattern of oocytedevelopment is asynchronous. Apart from the infer­ence that L. vittus is a multiple spawner, the numberof spawnings cannot be inferred from modes in thesize-frequency distribution lHunter & Goldberg 1980l.

Size-at-maturity

Ovaries were considered mature if their most advancedoocytes were yolked or ripe, as judged by the appear­ance of the whole oocyte. To avoid including maturefish that had regressed to an unyolked stage towardsthe end of the spawning season, only data taken be­tween September and February were used. The small-

est ripe female observed was 142 mm long (Fig. 4l. Allfish >220mm were mature (Fig. 5l. The Gompertz curvewas fit to the proportion of mature females by 10 mmlength groupings:

(ll

where Y(t) = proportion mature at length t, and y~, g,and to are parameters. This provided the best fit withthe least number of parameters. The length at which50% of females were mature (Lo.so) was calculated tobe 154mm. Lo.50 is reached at age-1 <Davis & West1992l, and all females at age-3 and older were mature(Fig. 5).

Spawning season

Only data on fish >200 mm were used to describe sea­sonal changes. The main spawning season appears tobe September-April (Fig. 6). The presence of fish withhydrating oocytes throughout the year indicates thespawning season is protracted, although only four fishwith hydrated oocytes were collected in June and Au­gust 1983. Some mature fish had only unyolked 00­

cytes in their ovaries during April-August (Fig. 6).The mean relative gonadal index lRGIl peaked in

September-October and then declined gradually, reach­ing its lowest value in June (Fig. 6), There was amarked and significant increase in mean RGI duringAugust-October (2x2 factorial ANOVA, month effect,F=309, df 1,339, p<O.OOl), with the same increase inboth years (interaction effect not significant: F=2.4,p>O.l), and there was a weak but significant differ­ence between years (F=4.3, p<0.05). There were sig­nificant differences in RGI between the four periodssampled during the main spawning season, Septem­ber 1982-April 1983 (ANOVA, F=50.5, df 3,706,p<O.OOl l, predominantly due to a linear componentW=130.1, df 1,706, p<O.OOl) caused by a decline inRGI over this period (slope -0.236, SE 0.021l. On theother hand, there was no significant difference in pro­portion of ripe fish throughout this period (likelihoodratio x2 = 6.3, df 3, p=O.lOl, nor any indication of adecline in the proportion. This suggests that spawningactivity remained at about the same level during themain spawning period but that gonads of individualfish were depleted by successive spawnings.

Lunar periodicity in spawning activity

Ovaries of mature fish (11.=374) sampled in November­December 1982 were examined histologically for evi­dence of spawning activity. Data were grouped into 1dperiods according to moon age. Two measures of spawn­ing were used: proportion of fish with postovulatory

228 Fishery Bulletin 91 (2), J993

Ripe

y =A+B sin Ix) +C cos (x) + D sin (2x) +

E cos (2x), (2)

where t=th9 x 21t, and t = moonage (d). A regression weighted bythe number (n) in each samplewas fit to arcsine (angular) root­transformed data. Proportions of1 were replaced by n-1/4 dividedby n. The fitted model accountedfor 85.9% of variance in the data.The smaller peak occurred 3d af­ter the new moon, and the largerpeak 6d after the full moon.

No simple pattern of spawn­ing was apparent from the pro­portion of fish with late migra­tory nucleus or hydrated oocytestages captured each day, pre­sumably because such a patternwas confounded by the effects oftime of day Isee next section).As postovulatory follicles prob­ably persist for a day, and maybelonger in other species (Hunter& Goldberg 1980, Hunter et al.1986), their detection does notdepend on the time of day ofsampling.

follicles, and proportion of fishwith late-migratory nucleus orhydrated-oocyte stages. Theformer showed a clear cyclicalpattern with two peaks of spawn­ing activity during the month ofsampling (Fig. 7). Various sinu­soidal curves were fit to the pro­portion of mature fish withpostovulatory follicles sampledeach day. The following model,which has a period of 29 d andallows one or two peaks of possi­bly unequal heights per period,was chosen:

700 BOO 900 1000

• hydrated (translucent)• late migratory nucleus (becoming translucent)• yolk granule (opaque)~ yolk vesicle (translucent)o perinucleolus (transparent)

Yolked

Ripe

Yolked

Unyolked

?

100 200 300 400 SOOo

5

5

o10

60

40

20

030

20

10

020

~>-Ul::CD 10~

tT~U.

010

Oocyte diameter (Jlm)

Figure 3Oocyte size-frequency distribution and oocyte stage. by 20l1m intervals. in Lutjanus{littu.~ ovaries representing the developmental sequence of maturation (see Table 1 fordetails). An ovary classified as Ripe indicates that it contained late-migratory nucleus­stage oocytes or ripe oocytes. Only oocytes >20011m diameter were measured in Ripestage-5 ovaries, whereas in other ovaries all oocytes >10011m were measured.

Die. periodicity inspawning activity

The November-December 1982samples were also examined forevidence of diel periodicity inspawning. Only mature fish

Davis and West: Reproductive biology of LUljanus vittus from North West Shelf of Australia 229

Figure 5Proportion of total female Lutjanus !littus that are mature by10 mm length-classes and by age-class during the height of thespawning season, Shown are 95% binomial confidence limits.Age data from Davis & West (1992).

The proportion of fish with early- or late-stage post­ovulatory follicles differed significantly with time ofsampling (early-stage likelihood ratio x2=130. df 7.p<0.001; late-stage likelihood ratio x 2=131, df 7,p<0.001 I. Fish with early-stage postovulatory follicleswere first detected at 12:00 h. By 17:00 h the propor­tion of fish with early-stage postovulatory follicles hadreached its highest level (91%l. Thereafter the propor­tion declined, until by 04:00 h none were present infish sampled. Late-stage postovulatory follicles followeda similar temporal pattern, with a peak that laggedthe early stage by 12-14h. The very few late-stagepostovulatory follicles in samples at 17:00h may haveresulted from spawning early that day or late spawn-

o

11TT

t

2 3 4 5 6Age (years)

I I I I I

150 200 250 300 350Length (mm)

I0.2

0.0

1.0

0.8

0.0100

1.0

l!?::J

~ 0.6

0.8

l!?::J1ii 0.6Es:::o:e 0.48.eQ. 0.2

s:::oi 0.4eQ.

•

o ripe• yolked• unyolked

•••

o1200

1000

- 800E~Cl 6000~

400

200

0100

Figure 4Diameter of the largest oocyte (MOD) of Lutjanus vittusby fish length. Ovaries classified as Yolked and Ripe indi­cate that fish are mature. Data were collected at height ofthe spawning season (October-February),

caught during days of major spawning activity (lu­nar days 2-10 and 18-29) were considered. Twomeasures that proved useful in detecting changesin spawning activity on temporal scales of less thana day were the proportion of mature fish with ripe­stage ovaries based on whole-oocyte staging, andmaximum oocyte diameter (MOD).

A clear diel cycle of spawning was evident(Fig. 8). Proportions of ripe-stage fish in samplestaken at different times throughout the day weresignificantly different (likelihood ratio x2=69. df 7,p<O.OOll. Proportions of ripe fish were highest be­tween 08:00 and 14:00 h. and no ripe fish werepresent by 16:00h. The mixture of ripe and unripeovaries between 11:00 and 15:00 h could indicatethat spawning for that day had already begun andthat some of the fish were spent or that only aportion of fish spawned each day. The temporal dis­tribution of MOD showed a similar pattern. Fishabout to spawn that day were clearly separable fromother fish by MOD at 11:00h. The MOD's of all butone fish sampled after 15:00 h were the same asnonspawning fish. suggesting that most spawningoccurred between 11:00 and 15:00 h on these days,which more or less coincided with rising daytimetides.

Postovulatory follicle data from the same subsetof fish also showed a diel pattern (Fig. 9. page 232).

200 300 400Length (mm)

230 Fishery Bulletin 91 (2). J 993

(Fig. 11)

Batch fecundity

Spawning frequency

As postovulatory follicles in L.vittus persist for -24 h. they pro­vide a good indication of spawn­ing over the previous 24 handcan be used to determine spawn­ing frequency. Parameter A fromEq. 2 provides an estimate of theproportion of females spawningeach day averaged over the 29 dlunar cycle. On average. femalesspawn 21.7 times each lunarcycle (95%CL=19.3-23.9>. If weassume that this spawning inten­sity occurs during October-April,then females spawn at least 150times each year.

...

67

S 0A

and the relation between fish total weight (W, g) andbatch fecundity by the equation

1In order to determine the mostappropriate sites in the ovaryfrom which to take subsamplesto determine batch fecundity, thenumber of eggs/mg tissue wasdetermined for six regions of theovary of five fish. Wedges of ovarywere taken from both inner andouter walls at anterior, middle,and posterior parts of each right­hand ovary. There were no sig­nificant differences in numbersof eggs between inside and out-side (ANOVA. F=0.016, df 1,4,p=0.91) or anterior, middle. orcenter (ANOVA. F=1.45. df 2,8,

p=0.29) nor any interaction between the effects (ANOVA.F=0.54, df 2.8, p=0.60). A paired t-test indicated thategg counts from a wedge of tissue taken from the middle­outside position did not differ significantly between leftand right ovaries (t=-0.008, df4,p=0.99). Subsequently,batch fecundity was estimated from the mean of threesubsamples taken from the anterior, middle, and poste­rior of either the left or right ovary of 29 fish. Thecoefficient of variation for estimates from threesubsamples was 0.02-0.21, with a mean ofO.lI.

The relation between fish length (L. mm) and batchfecundity (F) was best described by the equation

F = 3.656 X 10-6 U093 (F=26.1,df 1,27,p<0.001),

D ripe

176

• yolked

365

93

• unyolked

14074

L +53

+A S 0 N D J F M A M J J.. 1982 .. ... 1983

O-+----.-..------r----.-..-----r---.--,....---r---.--,....---r---.--,-----,

o

50

20

40

60

80

100

150

100

Figure 6Mean relative gonadal index, 95% confidence limits and range. and proportion of femaleLutjanus l,ittus at each maturity stage determined by whole-oocyte staging plotted againstthe mean date of each sampling period. Number in each sample is shown.

ing on the previous afternoon. In either case, thelate stage appears to persist for about 24 h. Only 13%of fish examined from this subset did not havepostovulatory follicles, and most of these (74%) werefrom fish sampled between 09:00 and 15:00 h. Thiswould be consistent with the hypotheses that folliclespersist for 24 h or less and that most fish spawn dailyduring the major periods of spawning activity.

The diel pattern of spawning indicated by data col­lected in November-December 1982 was supported byfurther sampling every 4 hover 6-9 October 1988. Theproportions of ripe fish in samples taken at differenttimes throughout the day were significantly differentIlikelihood ratio x2=1I0, df 6, p<O.OOl). Ripe fish werefound between 12:00 and 16:00h (Fig. 10>. The tempo­ral distribution of MOD followed the same pattern ob­served in November-December 1982.

Davis and West: Reproductive biology of LUljanus vittus from North West Shelf of Australia 231

3

III. , I, Ili.

3

I

0.8

0.6

0.4

0.0

0.2

400

600

800

1000

1200

~

1.0

~'C

Co'E8.~a..

Discussion

200

o 4:00 8:00 12:00 16:00 20:00 24:00

Time of day

Figure 8Proportion of mature female Lufjanlls vittus with ripe-stageovaries. and maximum oocyte diameter (MOm in individualscaught at different times of the day. Samples taken duringperiods of major spawning activity (lunar days 2-10 and 18­291 during November-December 1982. Oocytes have been as­signed to ovarian maturity stages: Ripe (0) and Yolked Ie>.Shown are 95% binomial confidence limits for proportionsand number in each sample.

The smallest ripe female observed in this study was142 mm long, which is close to the length-at-first­maturity of female L. vittus from New Caledonia(Loubens 1980a). In a review of size-at-first-maturity,Grimes (1987) found consistent differences between in­sular and continental species of lutjanids and betweenshallow and deepwater groups in the ratio of length­at-first-maturity to maximum length. Female L. vittusclearly fall into the continental shallow group that ma­ture early at -44% of their maximum length.

The spawning season on the NW Shelf is pro­tracted: While L. vittus spawned throughout the year.the major spawning period appeared to be September­April. Loubens (1980al determined from GSI andvisual staging that L. vittus spawn during October­February in New Caledonia (Loubens 1980al. Lutjanus

8 12 16 20 24 28

Moon age (days)

I I I I I Iii i I I I i I I I I I I I I i I I i I I I I I

tnew

9.

29 4

~j I II111I11

11 II1II

I11

o """"""'""""""""'_ 90

~ 80g 70

:§ 60aE 50Ql

.r. 40

'i 30c::~ 208.e 10a.. 0

F = 124.2 X W-3081 tF=17.7, dfl,27,p<0.0011.

Batch fecundity varied markedly between fish of thesame size. As the data were obtained from four differ­ent periods during the spawning season, we tested tosee whether the different spawning periods could ex­plain some of the variability in batch fecundity. Thereis some suggestion that batch fecundity might be higherin November and decline over the spawning season,following the seasonal pattern of RGI (Fig. 11.1. How­ever. in the regression of log fecundity/log length. thetest for homogeneity of slopes between sampling peri­ods was found to be not significant (ANCOVA, F=0.318,df 1,2604, p=0.57 I, and. assuming a common slope,there was no significant difference in the interceptsfor the four periods (ANCOVA, F=1. 76, df 1,2605,p=0.191.

Oven-dried weights of hydrated oocytes were deter­mined on 9 fish. Egg weight did not vary with fishlength (F=O.28, df 1.8. p=0.61) or with maximum oo­cyte diameter (F=O.88, df 1,8, p=O.37). The mean dryweight of individual hydrated oocytes was 0.012 mg(SE 0.0005).

Figure 7Proportion of mature female Lutjanus l'if/us with early­or late-stage postovulatory follicles plotted against time ofsampling (expressed as moon agel. Asinusoidal curve witha period of 29 d and two peaks of unequal height were fitto arcsine Iangular) root-transformed data. Lunar tidecycle, indicated by maximum and minimum heights ofdaily tides, has also been plotted. and the number in eachsample is shown.

232 Fishery Bulletin 9 J (2). 1993

1.0

0.0

32

1

27

l

6

13

157

..L

0.4

0.6

0.2

0.8

0.0

I :::~ (al>.

m0.6jc: 0.4~&.2 0.2

Q.

600

400

i•I Io

••••II

800

200

E 1200~..~ 1000

~'6

iE::;,E

~

31.0 -IIIg:§ 0.8CD]l.c 0.6'i5 0.4'E&.£ 0.2

Figure 9Proportion of mature female Lutjanus vittus with (a) earlyand Cb) late-stage postovulatory follicles plotted againsttime of sampling. Samples taken during periods of spawn­ing (lunar days 2-10 and 18-29) during November-De­cember 1982. Shown are 95% binomial confidence limitsfor proportions and number in each sample.

Figure 10Proportion of mature female Lutjanus vittus with ripe ova­ries. and maximum oocyte diameter (MODI of individualscaught at different times of the day. Samples taken during 6­10 October 1988. Oocytes have been assigned to ovarian ma­turity stages: Ripe (CII and Yolked (.1. Shown are 95%binomial confidence limits for proportions and number in eachsample.

0.0

o 4:00 8:00 12:00 16:00 20:00 24:00

Time of day

o 4:00 8:00 12:00 16:00 20:00 24:00

Time of day

vittus tends to comply with Grimes'119871 generaliza­tion that continental species, regardless of latitude,have a restricted spawning season.

The pattern of restricted spawning in L. vittus maybe linked with the production cycle on the NW Shelf,as has been suggested for other continental species oflutjanids (Grimes 1987). The nutrient source is slope­water washing up onto the NW Shelf in summer whenthe Leeuwin Current is no longer flowing (Holloway eta1. 1985, Tranter & Leech 1987). Enrichment is great­est between December and April. In winter the south­east trade winds blow, there is little stratification ofthe water column, and the plankton are dispersed(Tranter & Leech 1987 I. These winds abate by lateAugust or early September, enabling the water col­umn to become highly stratified and the plankton moreconcentrated (Tranter & Leech 19871. This would re­sult in improved feeding conditions for larvae and co­incides with the start of major spawning activity.

During the major spawning period, individual L.vittus spawn a number of times. Serial spawning has

been inferred in a number of lutjanids: L. purpureus(De Moraes 1970), L. kasmira IRangarajan 1971), L.griseus (Campos & Bashirullah 1975), L. synagris(Erhardt 1977), Pristipomoides multidens and P. typus(Min et al. 19771, Rhomboplites aurorubens (GrimesHuntsman 1980), P. filamentosus (Ralston 1981), Eteliscarbunculus (Everson 19841, E. corltscans and Aprionvirescens (Everson et al. 19891. While serial spawningappears to be commonplace in lutjanids, the numberof batches of eggs spawned each season has not beendetermined conclusively for any species (Grimes 1987).Our data suggest that L. vittus spawn about 22 times/mo during late November and early December. If thisspawning intensity were maintained throughout thewhole spawning period (October-April), then most in­dividuals would spawn about 150 times/yr. Even ifspawning intensity were half this rate for the remain­der of the season, then L. vittus would spawn about 90times/yr.

Greatest spawning activity in L. vittus was shortlyafter the full and new moons. A lunar rhythm in spawn-

Davis and West: Reproductive biology of Luljanus vittus from North West Shelf of Australia 233

<1 d. whereas studies on other species have found thatthey persist for <15 h (Takita et al. 1983). <2 d (Hunter& Goldberg 1980, Hunter et al. 1986, Clarke 1987), orslightly longer (Alheit et al. 1984). Our data onpostovulatory follicles suggest that females on averagespawn about 22 times each lunar month. The propor­tion of females spawning each day peaks about 3 dafter the new moon and 6 d after the full moon.

It seems unusual that L. vittus spawns during earlyafternoon. Thresher (1984) reviewed 22 pelagic egg­producing families. including lutjanids, and found thatmost spawned around dusk, as did Colin & Clavijo0988> who examined 26 species associated with coralreefs. Lutjanids produce pelagic eggs (Suzuki & Hioki1979> and all other reports indicate that they spawnat or after dusk (Wicklund 1969, Starck 1970, Suzuki& Hioki 1979). Possibly spawning is linked to the stateof the tide, rather than the time of day per se. In both1982 and 1988, spawning activity coincided with ris­ing tides; it may have been coincidental that thesewere during daylight hours.

Batch fecundity increased exponentially with fishlength and linearly with fish weight. The variability inbatch fecundities did not appear to be due to errors inusing the gravimetric method. The mean coefficient ofvariation of estimates was 0.11. and it never exceeded0.21. Care was taken to exclude ovaries that had par­tially spawned (identified by the presence of early-stagepostovulatory follicles), and only those ovaries that hada clearly identified mode of partly-hydrated oocyteswere used for estimates. Also. batch fecundities didnot appear to differ between sampling periods. Onemight have expected a decline in fecundity as ovaryweight declined over the season. Another possible rea­son for the variability in batch fecundity might havebeen that samples were taken at different phases ofthe lunar cycle. However, our data are inadequate totest this hypothesis.

Comparison of the fecundity of L. vittus with that ofother lutjanids is not particularly useful, due to incon­sistencies in methods of estimation and uncertainty asto whether true batch fecundities have been estimated(see review by Grimes 1987>. It is, however, useful toconsider the energetic costs of spawning by L. vittusconsidering its high annual fecundity. A 300 mm longfemale L. vittus would shed about 7.6 million eggsannually if it spawned 150 times/yr, or 4.5 million eggsif it spawned 90 times/yr. Assuming a calorific contentfor dried eggs of 25 kJ/g (the mean for a range of ga­doids; Hislop & Bell 1987), then the annual energyexpended on spawning (not including reproductive be­havior) would be -1300kJ for 90 spawnings, and2200 kJ for 150 spawnings. Estimation of reproductiveeffort or efficiency (egg energy/dietary energy x 100)

320

o September0

0

0 November 0 0

• February 0

0

• April0 •

8 00 0

0

• • 0

• •

• 0

•0

• • •

0

4 ,r----...,,------.-,-----.-,_----,160 200 240 280

Length (mm)

Figure 11Relation between batch fecundity and fish length of Lutjanusuittus at four different spawning periods (scales are logged I.

M

~ 40xenC>

~"0 20

~-g,.c:

~ 10I-

80

60

ing has been reported in many teleosts and somelutjanids, with most spawning activity close to the timeof the full moon; e.g., L. vaigiensis lRandall & Brock1960). L. griseus (Starck 1970). L. synagris (Reshet­nikov & Claro 1976), L. kasmira, L. rufolineatus, andP. multidens (Mizenko 1984; although aquarium popu­lations of L. kasmira spawned at the new and fullmoon [Suzuki & Hioki 1979]),

Many hypotheses have been proposed to explain whylunar reproductive cycles have developed in fishes (seereviews by Johannes 1978. Thresher 1984, Gladstone& Westoby 1988, Robertson et al. 1990). We cannotadequately explain why L. vittus has a well-developedlunar reproductive pattern. Spawning is probablylinked to the lunar tidal cycle rather than moonlightper se, as it occurred chiefly after new and full moonsand during daylight. The North West Shelf is a regionof strong semidiurnal tides which flow predominantlycross-shelf at the shelf break and alongshelf near thecoast (Holloway 1983). Spawning during spring tideswould therefore disperse the eggs ofL. vittus alongshelfrather than cross-shelf. There would be a clear advan­tage for larvae if they were advected to the midshelf orfurther offshore because these areas have greater pro­ductivity (Young et al. 1986). There is, however. noevidence for dispersal offshore; both larvae and adultsof Lutjanus spp. are primarily found on the inner shelf(Leis 1987 I.

Frequency of spawning was determined from theproportion of females with postovulatory follicles. Ourfindings indicate that postovulatory follicles persist for

234

requires information on daily ration, which is not avail­able for L. vittus and apparently has not been deter­mined for any other lutjanid. Nemipterous furcosis, aco-occurring species on the N.W. Shelf. has a dailyration in the wild of 2.3% for 23-29 em fish (KJ.Sainsbury, CSIRO Div. Fish., pers. commun. 19921.Epinephalus guttatus of similar size to L. vittus have adaily ration of 4% when fed surplus food once each dayin the laboratory (Menzel 1960). Assuming a some­what generous daily ration of 4% for L. vittus and amean prey energy content of 4.2 kJ Ig wet weight (Crisp1971) would result in an approximate annual repro­ductive efficiency of 4.8% for 90 spawnings and 8% for150 spawnings. This is similar to an annual reproduc­tive effort in Engraulis mordax of 8-11% lHunter &Leong 1981), a much smaller, shorter-lived species. Thereproductive efficiency of the goby Pomatoschistusmicrops is considered to be among the highest evercalculated (Rogers 1988), although its efficiency wascalculated using energy consumed over the 16 wk pe­riod of spawning and not the whole year. The reproduc­tive efficiency of P. microps, calculated using annualenergy intake, would be -8.7-13.5%. The reproductiveefficiency of L. vittus is comparatively high consideringthat this effort is sustained over many years.

AcknowledgmentsWe thank P. Binni and D. Le for laboratory assistance.and all the people who assisted in the fieldwork on theNorth West Shelf Program. K Haskard provided sta­tistical advice and modeled lunar periodicity in spawn­ing activity. S. Blaber, J.S. Gunn, R.E. Johannes, andtwo anonymous referees reviewed the manuscript andsuggested many improvements.

CitationsAlbeit, J., v: H. Alarcon, & B. J. Macewicz

1984 Spawning frequency and sex ratio in the Peruviananchovy, Engraulis ringens. Calif. Coop. OceanicFish. Invest. Rep. 25:43-52.

Campos, A. G., & A. K. M. Bashirullah1975 Biologia del parga Lutjanus griseus (Linn.) de la

Isla de Cubagua, Venezuela. II. Maduraci6n sexualy fecundidad. Bol. Inst. Oceanogr. Univ. OrienteCumana 14:109-116.

Clarke, T. A.1987 Fecundity and spawning frequency of the Hawai­

ian anchovy or nehu, Encrasicholina purpurea. Fish.Bull., U.S. 85:127-138.

Colin, P.L., & I.E. Clavijo1988 Spawning activity of fishes producing pelagic eggs

on a shelf edge coral reef, southwestern PuertoRico. Bull. Mar. Sci. 43:249-279.

Crisp, D. J.1971 Energy flow measurements. In Holmes, N.A., &

Fishery Bulletin 91 (2). J993

A.D. McIntyre (eds.), Methods for the study of ma­rine benthos, p. 197-323. IBP (Int. BioI. Prog. IHandb.16.

Davis, T. L. 0., & G. J. West1992 Growth and mortality of Lutjunus lJittus (Quoy

and Gaimardl from the North West Shelf ofAustralia. Fish. Bull., U.S. 90:395-404.

De Moraes, N. U.A.1970 Sobre a desova e a fecundidade do pargo, Lutjanus

purpureus Poey no nordeste Brasiliero. Bol. Estud.Pesca 10(1):7-20.

Erhardt, H.1977 Beitrage zur biologie von Lutjanus synagris

(Linnaeus 1758) an der KolumbianischenAtlantikkuste. Int. Rev. Gesamten Hydrobiol.62:161-171.

Erikson, D. L., Hightower, J. E., and G. D. Grossman1985 The relative gonadal index: an alternative index

for quantification of reproductive condition. CompoBiochem. Physiol. 81A (1):117-120.

Everson, A. R.1984 Spawning and gonadal maturation of the ehu.

Etelis carbunculus, in the northwestern HawaiianIslands. In Grigg, RW.• & K.Y. Tanoue (eds.), Proc.,Second symposium on resource investigations in theNorthwestern Hawaiian Islands, vol. 2, p. 128­148. UNIHI-SEAGRANT-MR-84-01, Univ. Hawaii,Honolulu.

Everson, A. R., H. A. Williams, & B. M. Ito1989 Maturation and reproduction in two Hawaiian

eteline snappers, uku, Aprion virescens, and onaga,Etelis coruscans. Fish. Bull., U.S. 87:877-888.

Forberg, K.G.1983 Maturity classification and growth of capelin. Mal­

lotus villosus (M.I. oocytes. J. Fish. BioI. 22:485-496.Gladstone, W., & M. Westoby

1988 Growth and reproduction in Canthigaster valentini(Pisces. Tetraodontidael: a comparison of a toxic reeffish with other reef fishes. Environ. BioI. Fishes21:207-221.

Grimes, C. B.1987 Reproductive biology of the Lutjanidae: a re­

view. In Polovina, J.J., & S. Ralston (eds.), Tropicalsnappers and groupers: biology and fisheries man­agement, p. 239-294. Westview Press, Boulder.

Grimes, C. B., & G. R. Huntsman1980 Reproductive biology of the vermilion snapper,

Rhomboplites aurorubens (Cuvier), from North andSouth Carolina waters. Fish. Bull., U.S. 78:137-146.

Hilge, 'Yo1977 On the determination of the stage of gonad ripe­

ness in female bony fishes. Meeresforschung 25:149­155.

Hislop, J. R. G., & M. A. Bell1987 Observations on the size, dry weight and energy

content of the eggs of some demersal fish species fromBritish marine waters. J. Fish. BioI. 31:1-20.

Holloway, P. E.1983 Tides on the Australian North-West Shelf. Aust.

J. Mar. Freshwater Res. 34:213-230.

Davis and West: Reproductive biology of Lurjanus vittus from North West Shelf of Australia 235

Holloway, P. E., S. E. Humphries, M. Atkinson, &J.Imberger

1985 Mechanisms for nitrogen supply to the AustralianNorth-West Shelf. Aust. J.Mar. Freshwater Res.36:753-764.

Hunter, J. R., & S. R. Goldberg1980 Spawning incidence and batch fecundity in north­

ern anchovy, Engraulis mordax. Fish. Bull., U.S.77:641-652.

Hunter, J. R., & R. J. H. Leong1981 The spawning energetics of female northern an­

chovy. Engraulis mordax. Fish. Bull., U.S. 79:215­230.

Hunter, J. R., & B. J. Macewicz1980 Sexual maturity, batch fecundity, spawning fre­

quency, and temporal pattern of spawning for thenorthern anchovy, Engraulis mordax, during 1979spawning season. Calif. Coop. Oceanic Fish. Invest.Rep. 21:139-149.

Hunter, J. R., N. C. H. Lo, & R. J. H. Leong1985 Batch fecundity in multiple spawning fishes. In

Lasker, R. (ed.), An egg production method for esti­mating spawning biomass of pelagic fish: applicationto the northern anchovy, Engraulis mordax, p. 67­77. NOAA Tech. Rep. NMFS 36.

Hunter, J. R., B. J. Macewicz, & J. R. Sibert1986 The spawning frequency of skipjack tuna, Kat­

suwonus pelamis, from the South Pacific. Fish. Bull.,U.S. 84:895-903.

Jernakoff, P., & K. J. Sainsbury1990 CSIRO's northern demersal finfish stock assess­

ments: 1980 to 1990. Inf. pap. IP/6/90, Bur. RuralResour., Dep. Primary Ind. & Energy Australia, 169 p.

Johannes, R. E.1978 Reproductive strategies of coastal marine fishes in

the tropics. Environ. BioI. Fishes 3:65-84.Kimura, D. K.

1977 Statistical assessment of the age-length key. J.Fish. Res. Board Can. 34:317-324.

Leis, J. M.1987 Review of the early life history of tropical groupers

(Serranidael and snappers (Lutjanidael. In Polovina,J.J., & S. Ralston (eds. l, Tropical snappers andgroupers: biology and fisheries management. p. 189­237. Westview Press, Boulder.

Loubens,G.1980a Biologie de quelques especes de poissons du lagon

neo-caledonien. II. Sexualite et reproduction. Cah.Indo-Pac. II (1):41-72.

1980b Biologie de quelques especes de poissons du lagonneo-caledonien. III. Croissance. Cah. Indo-Pac. II(2l:101-154.

Menzel, D. W.1960 Utilization of food by a Bermuda reef fish Epine­

phalus guttatus. J. Cons. Perm. Int. Explor. Mer25:216-222.

Min, T. S. T., T. Senta, & S. Supongpan1977 Fisheries biology of Pristipomoides spp. (Family

Lutjanidae) in the South China Sea and its adjacentwaters. Singapore J. Primary Ind. 5(2):96-115.

Mizenko, D.1984The biology of the Western Samoan reef-slope snap­

per populations of: Lutjanus kasmira, Lutjallusrufolineatus, and Pristipomoid€s multidens. M.S. the­sis. School Oceanogr., Univ. Rhode Island, Kingston,66p.

Mori, K.1984 Early life history of Liltjanus vitta (Lutjanidael in

Yuya Bay, the Sea of Japan. Jpn. J. Ichthyol.30(4):374-392.

Ralston, S. V. D.1981A study of the Hawaiian deepsea handline fishery

with special refer!!nce to the population dynamicsof opakapaka, Pristipomoides filamentosus (Pisces:Lutjanidae l. Ph.D. diss., Univ. Wash., Seattle,204p.

Randall, J. E., & V. E. Brock1960 Observations on the ecology of epinephaline and

lutjanid fishes of the Society Islands, with emphasison food habits. Trans. Am. Fish. Soc. 89:9-16.

Rangaragan, K.1971 Maturity and spawning of the snapper. Lutjanus

kasmira (Forskal) from the Andaman Sea. Indian J.Fish. 18:114-125.

Reshetnikov, Y. S., & R. M. Claro1976 Cycles of biological processes in tropical fishes with

reference to Lutjanus synagris. J. Ichthyol. 16:711­723.

Robertson, D. R., C. W. Peterson, & J. D. Brawn1990 Lunar reproductive cycles of benthic-brooding reef

fishes: reflections of larval biology or adult biol­ogy? Ecol. Monogr. 60(3l:311-329.

Rogers, S. I.1988 Reproductive effort and efficiency in the female

common goby, Pomatoschistus microps (Kroyerl (Tel­eostei: Gobioideil. J. Fish. BioI. 33:109-119.

Schnute, J.1981 A versatile growth model with statistically stable

parameters. Can. J. Fish. Aquat. Sci. 38:1128-1140.Starck, W. A. II

1970 Biology of the gray snapper, Lutjanus griseus. InStarck, WA. II, & R.E. Schroeder (eds.), Investiga­tions on the gray snapper, Llltjanlls griseus, p. 1­150. Stud. Trop. Oceanogr. (Miami) 10.

Suzuki, K., & S. Hioki1979Spawning behavior, eggs and larvae of the lutjanid

fish, Lutjanus kasmira, in an aquarium. Jpn. J.Ichthyol. 26:161-166.

Takita, T., T. Iwamoto, S. Kai, & I. Sogabe1983 Maturation and spawning of the dragonet, Cal­

lionymus enneactis, in an aquarium. Jpn. J. Ichthyol.30:221-226.

Thresher, R. E.1984 Reproduction in reef fishes. T.F.H. Publ., Nep­

tune City NJ, 399 p.Tranter, D. J., & G. S. Leech

1987 Factors influencing the standing crop of phyto­plankton on the Australian Northwest Shelf seawardof the 40m isobath. Continental Shelf Res. 7(2l:115­133.

236

Wallace, R. A., & K. Selman1981 Cellular and dynamic aspects of oocyte growth in

teleosts. Am. Zool. 21:325-343.West, G.

1990Methods of assessing ovarian development in fishes:a review. Aust. J. Mar. Freshwater Res. 41:199-222.

Wicklund, R.1969 Observations on spawning of the lane snap­

per. Underwater Nat. 6:40.Yamamoto, K.

1956 Studies on the formation of fish eggs. Annual cycle

Fishery Bulletin 9 J(2), J993

in the development of ovarian eggs in the flounder.Liopsetta obscura. J. Fac. Sci. Hokkaido Univ. Ser.6, 12:362-373.

Young, P. C., & K. J. Sainsbury1985 CSIRO's North West Shelf program indicates

changes in fish populations. Aust. Fish. 44(3):16­20.

Young, P. C., J. M. Leis, & H. F. Hausfeld1986 Seasonal and spatial distribution of fish larvae in

waters over the North West Continental Shelfof West­ern Australia. Mar. Ecol. Prog. Ser. 31:209-222.